h2oai · pseudotensor · Mar 6, 2024 · Mar 6, 2024 · Mar 6, 2024 · Mar 6, 2024
diff --git a/README.md b/README.md
@@ -147,8 +147,12 @@ Note that for all platforms, some packages such as DocTR, Unstructured, BLIP, St
 ---
 
 #### macOS (CPU/M1/M2) with full document Q/A capability
-* One-click Installers (Experimental and subject to changes)
+* One-click Installers (Experimental and subject to changes, we haven't tested each and every feature with these installers, we encourage the community to try them and report any issues)
 
+  Mar 07, 2024
+  - [h2ogpt-osx-m1-cpu](https://h2o-release.s3.amazonaws.com/h2ogpt/Mar2024/h2ogpt-osx-m1-cpu)
+  - [h2ogpt-osx-m1-gpu](https://h2o-release.s3.amazonaws.com/h2ogpt/Mar2024/h2ogpt-osx-m1-gpu)
+
   Nov 08, 2023
   - [h2ogpt-osx-m1-cpu](https://h2o-release.s3.amazonaws.com/h2ogpt/Nov2023/h2ogpt-osx-m1-cpu)
   - [h2ogpt-osx-m1-gpu](https://h2o-release.s3.amazonaws.com/h2ogpt/Nov2023/h2ogpt-osx-m1-gpu)

diff --git a/dev_installers/mac/README.md b/dev_installers/mac/README.md
@@ -17,12 +17,31 @@ This document provide the details to build one click installers for MacOS. To ma
 
 - Clone `h2ogpt` from https://github.com/h2oai/h2ogpt.git
 - Create conda environment and installer all required dependencies, consult [build_mac_installer.sh](build_mac_installer.sh) for more details.
-- Run below commands to build the installer
+- Run below commands to build the spec file for installer, replace the `--name` appropriately depending on whether building for CPU only or with MPS (GPU) support
     ```shell
     cd h2ogpt
-    pyinstaller ./dev_installers/mac/mac_run_app.py -D -w --name=h2ogpt-osx-m1-cpu-debug --hiddenimport=h2ogpt --collect-all=h2ogpt --noconfirm --recursive-copy-metadata=transformers --collect-data=langchain --collect-data=gradio_client --collect-all=gradio --path=${CONDA_PREFIX}/python3.10/site-packages --collect-all=sentencepiece --add-data=./Tesseract-OCR:Tesseract-OCR --add-data=./poppler:poppler
+    pyi-makespec mac_run_app.py -F --name=h2ogpt-osx-m1-cpu \
+      --hidden-import=h2ogpt \
+      --collect-all=h2ogpt \
+      --recursive-copy-metadata=transformers \
+      --collect-data=langchain \
+      --collect-data=gradio_client \
+      --collect-all=gradio \
+      --collect-all=sentencepiece \
+      --collect-all=gradio_pdf \
+      --collect-all=llama_cpp \
+      --collect-all=tiktoken_ext \
+      --add-data=../../Tesseract-OCR:Tesseract-OCR \
+      --add-data=../../poppler:poppler
     ```
-
+- Edit the `h2ogpt-osx-m1-cpu.spec` and/or `h2ogpt-osx-m1-gpu.spec` and add below code block to `Analysis()`, to explicitly tell PyInstaller to collect all `.py` modules from listed dependencies.
+    ```
+    module_collection_mode={
+        'gradio' : 'py',
+        'gradio_pdf' : 'py',
+    },
+    ```
+- Run `pyinstaller h2ogpt-osx-m1-cpu.spec` to build the installer.
 ### Deployment Mode
 
 - Clone `h2ogpt` from https://github.com/h2oai/h2ogpt.git

diff --git a/dev_installers/mac/build_mac_installer.sh b/dev_installers/mac/build_mac_installer.sh
@@ -6,57 +6,82 @@ then
     echo "conda could not be found, need conda to continue!"
     exit 1
 fi
+
+# Remove old Tesseract and poppler deps
+rm -rf ./Tesseract-OCR poppler
+
 conda env remove -n h2ogpt-mac
 conda create -n h2ogpt-mac python=3.10 rust -y
 conda activate h2ogpt-mac
 
+pip install --upgrade pip
+python -m pip install --upgrade setuptools
+
 # Install required dependencies into conda environment
-pip install -r requirements.txt --extra-index https://download.pytorch.org/whl/cpu
+pip install -r requirements.txt --extra-index https://download.pytorch.org/whl/cpu -c reqs_optional/reqs_constraints.txt
 # Required for Doc Q/A: LangChain:
-pip install -r reqs_optional/requirements_optional_langchain.txt
+pip install -r reqs_optional/requirements_optional_langchain.txt -c reqs_optional/reqs_constraints.txt
+# Optional: PyMuPDF/ArXiv:
+pip install -r reqs_optional/requirements_optional_langchain.gpllike.txt -c reqs_optional/reqs_constraints.txt
+# Optional: Selenium/PlayWright:
+pip install -r reqs_optional/requirements_optional_langchain.urls.txt -c reqs_optional/reqs_constraints.txt
+# Optional: DocTR OCR:
+conda install weasyprint pygobject -c conda-forge -y
+pip install -r reqs_optional/requirements_optional_doctr.txt -c reqs_optional/reqs_constraints.txt
+# Optional: for supporting unstructured package
+python -m nltk.downloader all
 
 # Required for CPU: LLaMa/GPT4All:
 # For MPS support
 if [ -z "$BUILD_MPS" ]
 then
     echo "BUILD_MPS is not set, skipping MPS specific configs..."
+    pip uninstall llama-cpp-python -y
+    CMAKE_ARGS="-DLLAMA_METAL=off" FORCE_CMAKE=1 pip install -r reqs_optional/requirements_optional_llamacpp_gpt4all.txt -c reqs_optional/reqs_constraints.txt --no-cache-dir
 else
     if [ "$BUILD_MPS" = "1" ]
     then
         echo "BUILD_MPS is set to 1, running MPS specific configs..."
-        export CMAKE_ARGS=-DLLAMA_METAL=on  # remove if CPU MAC
-        export FORCE_CMAKE=1
+        pip uninstall llama-cpp-python -y
+        CMAKE_ARGS="-DLLAMA_METAL=on" FORCE_CMAKE=1 pip install -r reqs_optional/requirements_optional_llamacpp_gpt4all.txt -c reqs_optional/reqs_constraints.txt --no-cache-dir
     fi
 fi
-pip install -r reqs_optional/requirements_optional_llamacpp_gpt4all.txt --no-cache-dir
-
-# Optional: PyMuPDF/ArXiv:
-pip install -r reqs_optional/requirements_optional_langchain.gpllike.txt
-# Optional: Selenium/PlayWright:
-pip install -r reqs_optional/requirements_optional_langchain.urls.txt
-# Optional: for supporting unstructured package
-python -m nltk.downloader all
-# Addtional Requirements
-pip install https://h2o-release.s3.amazonaws.com/h2ogpt/chromamigdb-0.3.25-py3-none-any.whl
-pip install https://h2o-release.s3.amazonaws.com/h2ogpt/hnswmiglib-0.7.0.tgz
+pip install librosa -c reqs_optional/reqs_constraints.txt
 
 # Install PyInstaller
 pip install PyInstaller
 
 # Install and copy tesseract & poppler
-#brew install tesseract@5.3.3
-#brew install poppler@23.10.0
-cp -R /opt/homebrew/Cellar/poppler/23.10.0/ ./poppler
-cp -R /opt/homebrew/Cellar/tesseract/5.3.3/ ./Tesseract-OCR
-
+#brew install poppler
+#brew install tesseract
+cp -R /opt/homebrew/Cellar/poppler/24.02.0/ ./poppler
+cp -R /opt/homebrew/Cellar/tesseract/5.3.4_1/ ./Tesseract-OCR
 
 # Build and install h2ogpt
 make clean dist
 pip install ./dist/h2ogpt*.whl
 
 # Build Mac Installer
-# below command is used to build current .spec file replace it whenever use new configs
-# pyinstaller mac_run_app.py -F --name=h2ogpt-osx-m1-cpu --hiddenimport=h2ogpt --collect-all=h2ogpt --noconfirm --recursive-copy-metadata=transformers --collect-data=langchain --collect-data=gradio_client --collect-all=gradio  --collect-all=sentencepiece --add-data=./Tesseract-OCR:Tesseract-OCR --add-data=./poppler:poppler
+# below command is used to build current .spec file from project root, replace it whenever use new configs
+#pyi-makespec mac_run_app.py -F --name=h2ogpt-osx-m1-cpu \
+#  --hidden-import=h2ogpt \
+#  --collect-all=h2ogpt \
+#  --recursive-copy-metadata=transformers \
+#  --collect-data=langchain \
+#  --collect-data=gradio_client \
+#  --collect-all=gradio \
+#  --collect-all=sentencepiece \
+#  --collect-all=gradio_pdf \
+#  --collect-all=llama_cpp \
+#  --collect-all=tiktoken_ext \
+#  --add-data=../../Tesseract-OCR:Tesseract-OCR \
+#  --add-data=../../poppler:poppler
+
+# add below argument to Analysis() call in h2ogpt-osx-m1-cpu.spec file
+#module_collection_mode={
+#    'gradio' : 'py',
+#    'gradio_pdf' : 'py',
+#}
 if [ "$BUILD_MPS" = "1" ]
 then
     echo "BUILD_MPS is set to 1, building one click installer for MPS..."

diff --git a/dev_installers/mac/h2ogpt-osx-m1-cpu.spec b/dev_installers/mac/h2ogpt-osx-m1-cpu.spec
@@ -15,10 +15,16 @@ tmp_ret = collect_all('gradio')
 datas += tmp_ret[0]; binaries += tmp_ret[1]; hiddenimports += tmp_ret[2]
 tmp_ret = collect_all('sentencepiece')
 datas += tmp_ret[0]; binaries += tmp_ret[1]; hiddenimports += tmp_ret[2]
+tmp_ret = collect_all('gradio_pdf')
+datas += tmp_ret[0]; binaries += tmp_ret[1]; hiddenimports += tmp_ret[2]
+tmp_ret = collect_all('llama_cpp')
+datas += tmp_ret[0]; binaries += tmp_ret[1]; hiddenimports += tmp_ret[2]
+tmp_ret = collect_all('tiktoken_ext')
+datas += tmp_ret[0]; binaries += tmp_ret[1]; hiddenimports += tmp_ret[2]
 
 
 a = Analysis(
-    ['./mac_run_app.py'],
+    ['mac_run_app.py'],
     pathex=[],
     binaries=binaries,
     datas=datas,
@@ -28,6 +34,10 @@ a = Analysis(
     runtime_hooks=[],
     excludes=[],
     noarchive=False,
+    module_collection_mode={
+        'gradio' : 'py',
+        'gradio_pdf' : 'py',
+    },
 )
 pyz = PYZ(a.pure)
 

diff --git a/dev_installers/mac/h2ogpt-osx-m1-gpu.spec b/dev_installers/mac/h2ogpt-osx-m1-gpu.spec
@@ -15,10 +15,16 @@ tmp_ret = collect_all('gradio')
 datas += tmp_ret[0]; binaries += tmp_ret[1]; hiddenimports += tmp_ret[2]
 tmp_ret = collect_all('sentencepiece')
 datas += tmp_ret[0]; binaries += tmp_ret[1]; hiddenimports += tmp_ret[2]
+tmp_ret = collect_all('gradio_pdf')
+datas += tmp_ret[0]; binaries += tmp_ret[1]; hiddenimports += tmp_ret[2]
+tmp_ret = collect_all('llama_cpp')
+datas += tmp_ret[0]; binaries += tmp_ret[1]; hiddenimports += tmp_ret[2]
+tmp_ret = collect_all('tiktoken_ext')
+datas += tmp_ret[0]; binaries += tmp_ret[1]; hiddenimports += tmp_ret[2]
 
 
 a = Analysis(
-    ['./mac_run_app.py'],
+    ['mac_run_app.py'],
     pathex=[],
     binaries=binaries,
     datas=datas,
@@ -28,6 +34,10 @@ a = Analysis(
     runtime_hooks=[],
     excludes=[],
     noarchive=False,
+    module_collection_mode={
+        'gradio' : 'py',
+        'gradio_pdf' : 'py',
+    },
 )
 pyz = PYZ(a.pure)
 

diff --git a/src/vision/__init__.py b/src/vision/__init__.py